“Corporate open source strategy” where I work is just having a form that engineers can fill out to request to open source things, and a committee on the other end of the form to sign off. It’s similar to the process for speaking at a conference or publishing on the company blog. Management sometimes steers in the direction of more or less public content, but specific releases are always individual initiative by engineers who want to develop their project in the open. Tech brand wants our name associated with high quality work.

I've only managed small businesses not large ones, but personally I'd be fine 9 times out of 10 with a developer who asked to open source a project they had built an mvp/poc of, but that never got approved to be used at all.

I could even imagine approving of a policy for the open sourcing / licensing of code, where any code that's used or previously used by the company in any way needs to go through an approvals process if anyone wants to open source it, while anything created but never used has a much simpler barrier such as manager agreeing in writing that it's unneeded code and therefore eligible for instant open sourcing under a specific license and specific terms of release.

> "But this is different, this is about code written for and deployed by the company itself"

Written for, yes, but seemingly never deployed (except to the extent that it could be demo'd and rejected). From the article:

> [After looking at a product owned by an unrelated team in the company, he single-handedly decided to make what he thought would be a good add-on or sibling to it] "I demoed Box Sums to the Box Notes team at some point, and they nitpicked the UI and implementation details (“What if two people type in the same cell at the same time? They’ll just overwrite each other.” ). Nothing came of it, but I took the code and shoved it into my back pocket for a rainy day."

It's not impossible "nothing came of it" is a shortened version of "they said it seemed like an awesome tool but too far from the original scope to want to take on and commit to maintaining, and as they said there was no chance that decision would change my manager agreed to sign off on my releasing it under MIT license as is allowed for un-used code."

I do academic research and write a ton of code to support this. In grad school (many years ago) at a Midwestern state school, I try to release some code under GPL and get blocked by the school's tech transfer department. It's a program that was designed to support the lab research we were doing (LIMS, ordering, etc). It wasn't much, but it very much made our lab run better. In the end, they licensed it out to a start up that flamed out. The entire process was messy, but all I really wanted was to release it with a GPL license and get on with my work. That office made my life quite difficult through grad school.

Fast forward a few years and I'm now at Stanford and then later UCSF. I email the tech transfer office about some code I'm planning on publishing, expecting a similar back and forth. It took all of two minutes to get back an email:

Are you planning on making money with this code? If so, let us know. If not, any open source license is fine with us.

It was a quite refreshing change to deal with institutions that knew what they were doing w.r.t. IP.